1.1 引言

我们在修改网络模型，添加或删除模块，或者更改了某一层之后，直接加载原先的预训练权重，肯定是会报错的，因为原来的模型权重和修改后的模型权重之间的结构是不匹配的。

那么我们只想加载那些没有更改过的那个部分的权重来初始化，应该怎么做？

1.2 问题的产生

以ResNet34为例，我在原有模型基础上添加一个模块，以SEAttention为例：

然后，加载训练文件，可以看到，报错以下信息：

fc.0和fc.2缺少权重。因为我们之前的模型的预训练权重是没有这两个部分的。

1.3 解决方法一

问题所在：

我们先转去看训练文件：

在加载预训练权重时有这么一个函数"load_state_dict"

我们ctrl+p查看一下该函数的参数：

注意看，最后面有一个叫"strict"的参数，它的默认值是True,也就是说在默认值的情况下，如果预训练的权重和模型的权重关键字不一致，就会报错。所以我们需要把这里的strict的值更改为False。

也就是说，之前训练不能对应的关键字，直接选择忽视掉。相当于一个包容的关系。

我们此时再次运行：正常训练。

现在，我们需要去原来的预训练权重文件中去看，查看里面的字典关键字。

在模型文件中进行：

1.查看预训练权重的关键字：

#1.查看预训练权重的关键字：
pretrained_weights_path ="./resnet34-pre.pth"
state_dict = torch.load(pretrained_weights_path)
#获取预训练权重的关键字
pretrained_keys = state_dict.keys()
print("预训练权重的关键字：")
for key in pretrained_keys:
    print(key)

运行结果如下： 

预训练权重的关键字：
conv1.weight
bn1.running_mean
bn1.running_var
bn1.weight
bn1.bias
layer1.0.conv1.weight
layer1.0.bn1.running_mean
layer1.0.bn1.running_var
layer1.0.bn1.weight
layer1.0.bn1.bias
layer1.0.conv2.weight
layer1.0.bn2.running_mean
layer1.0.bn2.running_var
layer1.0.bn2.weight
layer1.0.bn2.bias
layer1.1.conv1.weight
layer1.1.bn1.running_mean
layer1.1.bn1.running_var
layer1.1.bn1.weight
layer1.1.bn1.bias
layer1.1.conv2.weight
layer1.1.bn2.running_mean
layer1.1.bn2.running_var
layer1.1.bn2.weight
layer1.1.bn2.bias
layer1.2.conv1.weight
layer1.2.bn1.running_mean
layer1.2.bn1.running_var
layer1.2.bn1.weight
layer1.2.bn1.bias
layer1.2.conv2.weight
layer1.2.bn2.running_mean
layer1.2.bn2.running_var
layer1.2.bn2.weight
layer1.2.bn2.bias
layer2.0.conv1.weight
layer2.0.bn1.running_mean
layer2.0.bn1.running_var
layer2.0.bn1.weight
layer2.0.bn1.bias
layer2.0.conv2.weight
layer2.0.bn2.running_mean
layer2.0.bn2.running_var
layer2.0.bn2.weight
layer2.0.bn2.bias
layer2.0.downsample.0.weight
layer2.0.downsample.1.running_mean
layer2.0.downsample.1.running_var
layer2.0.downsample.1.weight
layer2.0.downsample.1.bias
layer2.1.conv1.weight
layer2.1.bn1.running_mean
layer2.1.bn1.running_var
layer2.1.bn1.weight
layer2.1.bn1.bias
layer2.1.conv2.weight
layer2.1.bn2.running_mean
layer2.1.bn2.running_var
layer2.1.bn2.weight
layer2.1.bn2.bias
layer2.2.conv1.weight
layer2.2.bn1.running_mean
layer2.2.bn1.running_var
layer2.2.bn1.weight
layer2.2.bn1.bias
layer2.2.conv2.weight
layer2.2.bn2.running_mean
layer2.2.bn2.running_var
layer2.2.bn2.weight
layer2.2.bn2.bias
layer2.3.conv1.weight
layer2.3.bn1.running_mean
layer2.3.bn1.running_var
layer2.3.bn1.weight
layer2.3.bn1.bias
layer2.3.conv2.weight
layer2.3.bn2.running_mean
layer2.3.bn2.running_var
layer2.3.bn2.weight
layer2.3.bn2.bias
layer3.0.conv1.weight
layer3.0.bn1.running_mean
layer3.0.bn1.running_var
layer3.0.bn1.weight
layer3.0.bn1.bias
layer3.0.conv2.weight
layer3.0.bn2.running_mean
layer3.0.bn2.running_var
layer3.0.bn2.weight
layer3.0.bn2.bias
layer3.0.downsample.0.weight
layer3.0.downsample.1.running_mean
layer3.0.downsample.1.running_var
layer3.0.downsample.1.weight
layer3.0.downsample.1.bias
layer3.1.conv1.weight
layer3.1.bn1.running_mean
layer3.1.bn1.running_var
layer3.1.bn1.weight
layer3.1.bn1.bias
layer3.1.conv2.weight
layer3.1.bn2.running_mean
layer3.1.bn2.running_var
layer3.1.bn2.weight
layer3.1.bn2.bias
layer3.2.conv1.weight
layer3.2.bn1.running_mean
layer3.2.bn1.running_var
layer3.2.bn1.weight
layer3.2.bn1.bias
layer3.2.conv2.weight
layer3.2.bn2.running_mean
layer3.2.bn2.running_var
layer3.2.bn2.weight
layer3.2.bn2.bias
layer3.3.conv1.weight
layer3.3.bn1.running_mean
layer3.3.bn1.running_var
layer3.3.bn1.weight
layer3.3.bn1.bias
layer3.3.conv2.weight
layer3.3.bn2.running_mean
layer3.3.bn2.running_var
layer3.3.bn2.weight
layer3.3.bn2.bias
layer3.4.conv1.weight
layer3.4.bn1.running_mean
layer3.4.bn1.running_var
layer3.4.bn1.weight
layer3.4.bn1.bias
layer3.4.conv2.weight
layer3.4.bn2.running_mean
layer3.4.bn2.running_var
layer3.4.bn2.weight
layer3.4.bn2.bias
layer3.5.conv1.weight
layer3.5.bn1.running_mean
layer3.5.bn1.running_var
layer3.5.bn1.weight
layer3.5.bn1.bias
layer3.5.conv2.weight
layer3.5.bn2.running_mean
layer3.5.bn2.running_var
layer3.5.bn2.weight
layer3.5.bn2.bias
layer4.0.conv1.weight
layer4.0.bn1.running_mean
layer4.0.bn1.running_var
layer4.0.bn1.weight
layer4.0.bn1.bias
layer4.0.conv2.weight
layer4.0.bn2.running_mean
layer4.0.bn2.running_var
layer4.0.bn2.weight
layer4.0.bn2.bias
layer4.0.downsample.0.weight
layer4.0.downsample.1.running_mean
layer4.0.downsample.1.running_var
layer4.0.downsample.1.weight
layer4.0.downsample.1.bias
layer4.1.conv1.weight
layer4.1.bn1.running_mean
layer4.1.bn1.running_var
layer4.1.bn1.weight
layer4.1.bn1.bias
layer4.1.conv2.weight
layer4.1.bn2.running_mean
layer4.1.bn2.running_var
layer4.1.bn2.weight
layer4.1.bn2.bias
layer4.2.conv1.weight
layer4.2.bn1.running_mean
layer4.2.bn1.running_var
layer4.2.bn1.weight
layer4.2.bn1.bias
layer4.2.conv2.weight
layer4.2.bn2.running_mean
layer4.2.bn2.running_var
layer4.2.bn2.weight
layer4.2.bn2.bias
fc.weight
fc.bias

2.查看你自身网络模型的关键字：

#2.查看你自身网络模型的关键字：
net = resnet34()
model_keys = net.state_dict().keys()
print("\n模型权重的关键字：")
for key in model_keys:
    print(key)

模型权重的关键字：
conv1.weight
bn1.weight
bn1.bias
bn1.running_mean
bn1.running_var
bn1.num_batches_tracked
layer1.0.conv1.weight
layer1.0.bn1.weight
layer1.0.bn1.bias
layer1.0.bn1.running_mean
layer1.0.bn1.running_var
layer1.0.bn1.num_batches_tracked
layer1.0.conv2.weight
layer1.0.bn2.weight
layer1.0.bn2.bias
layer1.0.bn2.running_mean
layer1.0.bn2.running_var
layer1.0.bn2.num_batches_tracked
layer1.1.conv1.weight
layer1.1.bn1.weight
layer1.1.bn1.bias
layer1.1.bn1.running_mean
layer1.1.bn1.running_var
layer1.1.bn1.num_batches_tracked
layer1.1.conv2.weight
layer1.1.bn2.weight
layer1.1.bn2.bias
layer1.1.bn2.running_mean
layer1.1.bn2.running_var
layer1.1.bn2.num_batches_tracked
layer1.2.conv1.weight
layer1.2.bn1.weight
layer1.2.bn1.bias
layer1.2.bn1.running_mean
layer1.2.bn1.running_var
layer1.2.bn1.num_batches_tracked
layer1.2.conv2.weight
layer1.2.bn2.weight
layer1.2.bn2.bias
layer1.2.bn2.running_mean
layer1.2.bn2.running_var
layer1.2.bn2.num_batches_tracked
layer2.0.conv1.weight
layer2.0.bn1.weight
layer2.0.bn1.bias
layer2.0.bn1.running_mean
layer2.0.bn1.running_var
layer2.0.bn1.num_batches_tracked
layer2.0.conv2.weight
layer2.0.bn2.weight
layer2.0.bn2.bias
layer2.0.bn2.running_mean
layer2.0.bn2.running_var
layer2.0.bn2.num_batches_tracked
layer2.0.downsample.0.weight
layer2.0.downsample.1.weight
layer2.0.downsample.1.bias
layer2.0.downsample.1.running_mean
layer2.0.downsample.1.running_var
layer2.0.downsample.1.num_batches_tracked
layer2.1.conv1.weight
layer2.1.bn1.weight
layer2.1.bn1.bias
layer2.1.bn1.running_mean
layer2.1.bn1.running_var
layer2.1.bn1.num_batches_tracked
layer2.1.conv2.weight
layer2.1.bn2.weight
layer2.1.bn2.bias
layer2.1.bn2.running_mean
layer2.1.bn2.running_var
layer2.1.bn2.num_batches_tracked
layer2.2.conv1.weight
layer2.2.bn1.weight
layer2.2.bn1.bias
layer2.2.bn1.running_mean
layer2.2.bn1.running_var
layer2.2.bn1.num_batches_tracked
layer2.2.conv2.weight
layer2.2.bn2.weight
layer2.2.bn2.bias
layer2.2.bn2.running_mean
layer2.2.bn2.running_var
layer2.2.bn2.num_batches_tracked
layer2.3.conv1.weight
layer2.3.bn1.weight
layer2.3.bn1.bias
layer2.3.bn1.running_mean
layer2.3.bn1.running_var
layer2.3.bn1.num_batches_tracked
layer2.3.conv2.weight
layer2.3.bn2.weight
layer2.3.bn2.bias
layer2.3.bn2.running_mean
layer2.3.bn2.running_var
layer2.3.bn2.num_batches_tracked
layer3.0.conv1.weight
layer3.0.bn1.weight
layer3.0.bn1.bias
layer3.0.bn1.running_mean
layer3.0.bn1.running_var
layer3.0.bn1.num_batches_tracked
layer3.0.conv2.weight
layer3.0.bn2.weight
layer3.0.bn2.bias
layer3.0.bn2.running_mean
layer3.0.bn2.running_var
layer3.0.bn2.num_batches_tracked
layer3.0.downsample.0.weight
layer3.0.downsample.1.weight
layer3.0.downsample.1.bias
layer3.0.downsample.1.running_mean
layer3.0.downsample.1.running_var
layer3.0.downsample.1.num_batches_tracked
layer3.1.conv1.weight
layer3.1.bn1.weight
layer3.1.bn1.bias
layer3.1.bn1.running_mean
layer3.1.bn1.running_var
layer3.1.bn1.num_batches_tracked
layer3.1.conv2.weight
layer3.1.bn2.weight
layer3.1.bn2.bias
layer3.1.bn2.running_mean
layer3.1.bn2.running_var
layer3.1.bn2.num_batches_tracked
layer3.2.conv1.weight
layer3.2.bn1.weight
layer3.2.bn1.bias
layer3.2.bn1.running_mean
layer3.2.bn1.running_var
layer3.2.bn1.num_batches_tracked
layer3.2.conv2.weight
layer3.2.bn2.weight
layer3.2.bn2.bias
layer3.2.bn2.running_mean
layer3.2.bn2.running_var
layer3.2.bn2.num_batches_tracked
layer3.3.conv1.weight
layer3.3.bn1.weight
layer3.3.bn1.bias
layer3.3.bn1.running_mean
layer3.3.bn1.running_var
layer3.3.bn1.num_batches_tracked
layer3.3.conv2.weight
layer3.3.bn2.weight
layer3.3.bn2.bias
layer3.3.bn2.running_mean
layer3.3.bn2.running_var
layer3.3.bn2.num_batches_tracked
layer3.4.conv1.weight
layer3.4.bn1.weight
layer3.4.bn1.bias
layer3.4.bn1.running_mean
layer3.4.bn1.running_var
layer3.4.bn1.num_batches_tracked
layer3.4.conv2.weight
layer3.4.bn2.weight
layer3.4.bn2.bias
layer3.4.bn2.running_mean
layer3.4.bn2.running_var
layer3.4.bn2.num_batches_tracked
layer3.5.conv1.weight
layer3.5.bn1.weight
layer3.5.bn1.bias
layer3.5.bn1.running_mean
layer3.5.bn1.running_var
layer3.5.bn1.num_batches_tracked
layer3.5.conv2.weight
layer3.5.bn2.weight
layer3.5.bn2.bias
layer3.5.bn2.running_mean
layer3.5.bn2.running_var
layer3.5.bn2.num_batches_tracked
layer4.0.conv1.weight
layer4.0.bn1.weight
layer4.0.bn1.bias
layer4.0.bn1.running_mean
layer4.0.bn1.running_var
layer4.0.bn1.num_batches_tracked
layer4.0.conv2.weight
layer4.0.bn2.weight
layer4.0.bn2.bias
layer4.0.bn2.running_mean
layer4.0.bn2.running_var
layer4.0.bn2.num_batches_tracked
layer4.0.downsample.0.weight
layer4.0.downsample.1.weight
layer4.0.downsample.1.bias
layer4.0.downsample.1.running_mean
layer4.0.downsample.1.running_var
layer4.0.downsample.1.num_batches_tracked
layer4.1.conv1.weight
layer4.1.bn1.weight
layer4.1.bn1.bias
layer4.1.bn1.running_mean
layer4.1.bn1.running_var
layer4.1.bn1.num_batches_tracked
layer4.1.conv2.weight
layer4.1.bn2.weight
layer4.1.bn2.bias
layer4.1.bn2.running_mean
layer4.1.bn2.running_var
layer4.1.bn2.num_batches_tracked
layer4.2.conv1.weight
layer4.2.bn1.weight
layer4.2.bn1.bias
layer4.2.bn1.running_mean
layer4.2.bn1.running_var
layer4.2.bn1.num_batches_tracked
layer4.2.conv2.weight
layer4.2.bn2.weight
layer4.2.bn2.bias
layer4.2.bn2.running_mean
layer4.2.bn2.running_var
layer4.2.bn2.num_batches_tracked
fc.weight
fc.bias
se.fc.0.weight
se.fc.2.weight
 

3.找出模型中缺失/多余的权重

缺失

【针对你预训练关键字比模型关键字少】

#3.找出模型中缺失的权重【针对你预训练关键字比模型关键字少】
missing_keys = model_keys- pretrained_keys
print("\n模型中缺失的权重关键字")
for key in missing_keys:
    print(key)

运行结果如下： 

模型中缺失的权重关键字:
layer1.1.bn1.num_batches_tracked
layer2.1.bn1.num_batches_tracked
layer1.0.bn2.num_batches_tracked
layer4.0.downsample.1.num_batches_tracked
layer4.0.bn2.num_batches_tracked
layer3.5.bn1.num_batches_tracked
layer2.0.bn2.num_batches_tracked
layer3.2.bn2.num_batches_tracked
layer3.1.bn2.num_batches_tracked
layer2.2.bn2.num_batches_tracked
layer3.0.bn1.num_batches_tracked
layer4.1.bn2.num_batches_tracked
layer1.1.bn2.num_batches_tracked
se.fc.2.weight
layer4.2.bn1.num_batches_tracked
layer3.3.bn1.num_batches_tracked
layer3.4.bn2.num_batches_tracked
layer3.2.bn1.num_batches_tracked
layer2.3.bn1.num_batches_tracked
layer2.1.bn2.num_batches_tracked
layer3.0.downsample.1.num_batches_tracked
layer4.2.bn2.num_batches_tracked
layer2.2.bn1.num_batches_tracked
bn1.num_batches_tracked
se.fc.0.weight
layer1.2.bn2.num_batches_tracked
layer2.0.downsample.1.num_batches_tracked
layer3.1.bn1.num_batches_tracked
layer2.0.bn1.num_batches_tracked
layer3.5.bn2.num_batches_tracked
layer1.2.bn1.num_batches_tracked
layer4.0.bn1.num_batches_tracked
layer3.3.bn2.num_batches_tracked
layer1.0.bn1.num_batches_tracked
layer3.4.bn1.num_batches_tracked
layer3.0.bn2.num_batches_tracked
layer2.3.bn2.num_batches_tracked
layer4.1.bn1.num_batches_tracked

多余

【针对你预训练关键字比模型关键字多】

unexpected_keys = pretrained_keys -model_keys
print("\n预训练权重中多余的权重关键字：")
for key in unexpected_keys:
    print(key)

1.4 方法二：写一个判断语句

先找到训练文件的加载初始权重的代码：

  net = resnet34()
    # load pretrain weights
    # download url: https://download.pytorch.org/models/resnet34-333f7ec4.pth
    model_weight_path = "./resnet34-pre.pth"
    assert os.path.exists(model_weight_path), "file {} does not exist.".format(model_weight_path)
    net.load_state_dict(torch.load(model_weight_path, map_location='cpu'),strict=False)
    # for param in net.parameters():
    #     param.requires_grad = False

然后删掉，我们重新写一个：

#模板
    model_weight_path ="./resnet34-pre.pth" #预训练权重的路径
    ckpt = torch.load(model_weight_path) #加载预训练权重
    net = resnet34() #实例化模型
    model_dict = net.state_dict() #获取我们模型的参数

    #判断预训练模型中网络的模块是否在修改后的网络中也存在，并且shape相同，如果相同则取出
    pretrained_dict = {k:v for k,v in ckpt.items() if k in model_dict and(v.shape == model_dict[k].shape)}

    #更新修改之后的model_dict
    model_dict.update(pretrained_dict)
    
    #加载我们真正需要的state_dict
    net.load_state_dict(model_dict,strict=True)

该代码可作为模板使用。

可以看到同样可以运行：

1.5 多GPU训练的关键字问题

看着确实非常简单，但有时候会出现一个问题，比如：

有些同学在单GPU上调用一个基于多GPU预训练像这个问题，报错：

这是因为在加载多GPU训练的模型的时候，由于用DataParallel训练的模型数据并行方式的训练，key中会包含"module“关键字。去掉DataParallel预训练模型中的module，修改如下:

明显看出，我要加载的预训练权重和网络模型结构是一样，只是每个名字前面多了module.这几个后面是一模一样，如果我按照之前的说直接忽略，是不是每一个都对应不上，那就相当于没有加载，【主要原因就是模型的key不一致，也就是层的名称不一致比如我使用的是resnet模型，那么层名应该是conv1.weight而不是module.conv1.weight】

那么如何解决呢：

net就是你实例化的模型名字

checkpoint指的是前面提到的“ckpt”也就是预训练权重。

net.load_state_dict({k.replace('module.',''):v for k,v in checkpoint['state_dict'].items()})

附录：

在PyTorch中，使用torch.load()函数加载的预训练模型权重确实是一个字典类型，这个字典被称为状态字典（state_dict）。状态字典包含了模型中所有可学习参数（如权重和偏置）的键值对，键通常是参数的名称，值则是对应的Tensor（包含实际的数值）。

当你从文件中加载预训练权重时，代码通常看起来像这样：

pretrained_weights_path = 'path_to_pretrained_model.pth'
state_dict = torch.load(pretrained_weights_path)

这里的state_dict就是包含了模型权重的字典。之后，你可以通过model.load_state_dict(state_dict)来将这些权重加载到你的模型中，前提是你模型的结构与预训练模型的结构相匹配，或者你已经适当地处理了任何不匹配的情况。